Monday, October 31, 2016

eBEEF Monday: Economically Relevant Traits

Economically relevant traits (ERTs) are those that are directly associated with either a cost or a source of revenue.  Not all Expected Progeny Differences (EPDs) represent traits that are ERTs, and instead represent indicator traits. It is important for producers to know the difference between ERTs and indicator traits when making selection decisions.

For more information, see the eBEEF.org factsheet "Economically Relevant Traits."

Friday, October 28, 2016

Beefmaster Breeders United Convention: EPDs and Selection Indexes

Matt Spangler
University of Nebraska-Lincoln

In the past, the only way we made improvement was through visual appraisal.

Photo from Harlan Ritchie's Beef Review. Difference between 1835 and 1937 champion.
As the picture above shows, we can make a change, but how many of us want to wait 100 years?

Improvement can be accomplished through management and genetics.

In the past we (animal breeding scientist) have probably done a disservice to the industry by producing lots of EPDs, then dumping those in beef producers lap and then expect you all to make meaningful decisions with them. In some situations, this may be as valuable as a free cat.

There are many factors that can influence an animals record, for example a weaning weight. Weaning weight may be affected by age of the calf, age of the dam, how much it was feed, and other environmental factors.

So, we need to compare animals to their contemporaries, Contemporaries are animals of the same sex, raised at the same ranch with the same management. We can calculate a ratio of how different from the contemporary group each animal is. These ratios are helpful for making within herd selection, but it cannot be used to make selection decisions across herds. EPDs allow us to make selection decisions across herds.

What we need to be selecting on is the genetics, because this is inherited from generation and generation and allows us to make genetic progress.

Breeders often misunderstand EPD accuracy. Accuracy is not a measurement of the consistency of the calf crop. Accuracy is a measure of how certain we are of the EPD estimate. Accuracy is really a measure of the possible change we might see in that EPD. The EPD may be 40, but the 68% confidence interval might be between 35 and 45 pounds.

Consider two bulls with a weaning weight EPD of 40. The first bull has an EPD ACC of 0.3, which means a possible change 8.1 pounds. The second bull has an EPD ACC = 0.8, which means a possible change of about 2 pounds.

Another very important tool is percentile rank. If a bull is in the 25th percentile, he is better than 75% of the other bulls in the breed and worse than 24% of the bulls in the breed.

The across-breed EPD adjustment factors are a tool to compare EPDs across different breeds. By adding or subtracting by the adjustment factors.

There is no easily accessible, objective way for breeders, particularly breeders in the beef and sheep industries where ownership is diverse and production environments vary a great deal, to use these predictions [EPDs] intelligently.  -- Bourdon
This is not a question of the intelligence of beef producers. This is commentary on the difficulty of using EPDs. The solution to this is to use economic selection indexes. Profit is of course revenue minus cost. It is quite easy for us to create EPDs for revenue traits. It is harder to create EPDs for the cost side.

Seedstock producers need to discuss the customers needs. Spangler shared the experience of seedstock producers calling him up to sell him bulls to use in the University of Nebraska's cattle herd. They start telling him about the bulls in the sale he should be looking at. These breeders have not asked him about his needs. Seedstock producers need to do a better job of asking their customers about their production. What is the customer's breeding objective? When are they marketing their calves? What is their production environment like?

We need to focus on economically relevant traits, and not indicator traits. Indicator traits are important to measure, record, and report because they help us predict economically relevant traits.
Lets consider pairs of traits. Which one is the economically relevant trait?
BWT vs CE: Birth weight is an indicator of calving ease. Calving ease can cause increased labor costs and leads to problems for cattle rebreeding.
REA vs YG: Ribeye area is a component of Yield Grade. Yield grade is the trait that influences the sale price, thus is the economically relevant trait.
YWT vs CWT: This comparison depends on when cattle are sold. But, for the overall industry, Carcass Weight is the economically relevant trait.
MWT vs DMI: Mature weight is an indicator of the dry matter intake of the cow. The dry matter intake is the economically relevant trait.
RFI vs FI: No one gets paid for residuals. You have to pay the entire feed bill. Feed intake, not residual feed intake, is the economically relevant trait.

There are different traits that are important depending on the production sector of the industry. If you are selecting for cattle

There are three methods for selecting for multiple traits.
First is tandem selection. In this strategy we select for birth weight till it reaches an appropriate level. Then we select for weaning weight. The problem with this strategy is that when we select for increase weaning weight we loose ground on the progress we made for birth weight (birth weight goes up), because the two traits are correlated. This is a very inefficient selection strategy!

Independent culling levels. In this method, you set minimum levels for birth weight, weaning weight, marbling, etc. This method is also inefficient because it fails to recognize bulls that are excellent for several traits but barely misses the cutoff for a trait. This also does not weight traits according to their economic importance.

Selection indices were first developed in 1943. The first selection index in the beef industry was published in 2004. This was a 60 year time lapse.
"This would be similar to corn producers hitching up the mules to plant the corn in the spring," Spangler said. This is a proven technology and needs to be adopted by the beef industry.

When designing a selection index we are trying to make improvement for traits in our breeding objective. The breeding objective may contain things like calf survival, weaning weight, fertility, etc. The more closely the traits for which there are EPDs match the traits in the breeding objective, the better the selection index works. When the breeding objective trait is not predicted, we need to use an indicator trait.

Why do we use the past five years of economic data and not the current prices and spreads? We are selecting cattle to use in the future, not today. A five year average is going to give us a better prediction of future prices than current prices will.

For the Beefmaster Terminal Economic Index, carcass weight is the most impactful trait, followed by feed intake and third by marbling.

For Beefmaster Maternal Economic Index, smaller mature weight is the most important trait, while increased weaning weight and increased maternal effect on growth (milk) are the second and third most importnat traits. This index selects bulls that will producer daughters that are smaller cows that wean heavier calves.

As Spangler and his group developed these indexes, they looked at how sensitive their indexes are to the assumed genetic correlations and the assumed economic values. Whether or not the calves were all calf feeds or yearling slaughter, had no effect on the index.
They found that the indexes were very robust (insensitive) to changes to genetic correlations or economic assumptions.

Lets look at two bulls, one with an index of 100 and the second with an index of 76. Lets assume that over 4 years, these two bulls are exposed to 120 cows.
120 exposure X ($100 - $76) = $2880 in profit difference between the two bulls.

Improvement in current indices can be made by increasing the number of economically relevant traits that have EPD predictions.

  • Input traits
  • Fertility


Enterprise level profitability should move closer to industry level profitability. Cow-calf producers don't get paid for tenderness, but tenderness is a big driver of consumer demand.

Seedstock producers should focus on the indexes that influence their customer's profitability. If a seedstock producer's customers retain ownership, then the seedstock producer should focus on the terminal index. If the majority of the customers retain females and sell at weaning, then the seedstock producer should focus on the maternal index.

Progeny receive half of their genetic material from each parent.
Breeding Value = 1/2 Sire Breeding Value + 1/2 Dam Breeding Value + Theta.
What is theta? Theta is the Mendelian sampling term. This accounts for the random sample of genes a calf inherited from its sire and dam. Even really good bulls can produce bad calves due to this random shuffle of genes (see fact sheet).

Seedstock producers are not truly cattle producers, they are genetic providers. Genetic improvement, driven by accuracy, selection intensity, and generation interval, should be the focus of seedstock producers. Seedstock producers need to realize that using younger sires can decrease their generation interval and increase their genetic progress.

Keep in mind, that a genomic test increases accuracy. The accuracy will always increase, but the EPD estimate can go up or down. But, the incrase in accuracy can allow us to use younger bulls with more confidence.

GE-EPDs allow commercial producers more confidence that they are picking the right bull.

In the past, we have seen producers only testing what they thought were their top bulls. This is not optimum.
At UNL, they have space to put 90 bulls on feed. At weaning they genomically test every bull calf. With the genomic test, they now have EPDs on every trait. Before they were picking bulls somewhat blind. Now they have data and information behind which bulls go on feed.

"Genomics helps breeds the most that already have very sound databases," Spangler said.

One of the things that breed associations do poorly is collectively bargain for the price of technology. In Ireland they negotiated to genotyped all of the cattle in the country. Because they were able to buy one million SNP chips, the cost per tested animal was less than $20.


Wednesday, October 26, 2016

NBCEC Brown Bagger: Implementation of single step methodologies at Angus Genetics, Inc.

Steve Miller
AGI

Angus Genetic Services provides evaluations for AAA, CAA, and Charolais breed associations.

"The ship has sailed on using genomics. Breeders are using it now, and seeing the benefits of it," said Miller.

Previously at AGI, they have been using a two-step approach. In this method, a genomic prediction is created and then is used as an indicator trait for EPD estimation. The calibration data set size has increased dramatically as Angus breeders have used genomic-enhanced EPDs.

The orgininal method of incorporting genomic predictions as correlated trait.

In the future, we will stop referring to genomic-enhanced EPDs. We don't refer to EPDs as pedigree-enhanced or performance-enhanced, we simply refer to them as EPDs. In the future use of genomic data in genetic prediction will become so routine that we will simply call them EPDs.

Is the Animal Model Obsolute?


In single-step genomic prediction, we combine the measures of relatedness from pedigree data with the measures of relatedness from the genomic data. In comparison to pedigree data, genomic data captures more variation is relationships.

Consider 6 full sibs. Their pedigree relationship is 0.59 (slightly higher than 0.5 due to inbreeding in the pedigree). But with genomic data the relationships vary from 0.49 to 0.65.

Single-step
 eliminates the need for periodic calibration.
utilizes all available data

The migration to single-step genomic evaluation is not unique to beef cattle. This has happened in multiple breeds and species on multiple continents.

One of the keys that makes single-step genomic evaluation is APY.

The Angus association has 7.6 million birth weight records and 254,000 genotyped animals.

The move to single-step required more computing power at AGI; they have purchased 4 new servers.

Between the previous correlated evaluation and single-step evaluation, for  200 proven sires the correlation between the current and new EPDs were 0.99.

AGI is currently in the process of changing horses on the fly. They are currently running two genetic evaluations, the current two-step correlated genetic evaluation and the new single-step evaluation. They are looking at the consistency of these predictions over weekly runs.

Right now, when we recalibrate a genomic prediction, we can see big jumps in EPDs. But, with weekly evaluations as data is added, AGI sees  incremental changes in EPDs as phenotype data is added.

Single-step BLUP will allow progressive breeders to leverage all of their data. Phenotype and genotype information will be utilized together in a single step.

Angus Association has also started a sire progeny testing program to get carcass data on proven, popular sires.

The Angus Association has not yet identified a time to completely switch to single-step BLUP. They are evaluating single-step, and when they are confident that the program is ready, they will switch.

NBCEC Brown Bagger: Implementation of single step methodologies at International Genetic Solutions

Dr. Mahdi Saatchi
International Genetic Solutions

IGS performs genetic evaluations for 12 breed associations from North America.
IGS has over 16 million animals in their database and is adding over 400,000 animals per yer.
IGS has 84,197 animals with genotypes. Simmental makes up about 40,000 of these genotypes.

Currently at IGS they blend the molecular breeding value (MBV, the genomic prediction) with the multi-breed international cattle evaluation. This is more like the blending that occurs with selection indexes.

Single-step genomic prediction allows information from genotyped animals to be spread to related animals in the data set.
Also, multiple-step genomic predictions were often trained on breeding values, and any errors in the estimation of the breeding values influenced the genomic prediction.

There are two approaches to single-step genetic evaluation. Single-step BLUP uses a breeding value model. Single-step Bayesian Regression uses a marker effects model.

In Single-step Bayesian Regression does not requiring inverting the relatedness matrix.
Also, in single-step Bayesian regression allows us to give different importance to DNA markers used (i.e. variable selection). In the dairy industry, they observe little benefit from Bayesian regression models. But, in the beef industry, we see genomic regions (i.e. QTLs) that have large effects in multiple breeds. Bayesian regression allows us to fit large effects to certain DNA markers used in the genetic prediction.

In the single-step Bayesian regression model, we infer (in other words predicting or imputing) the genotypes for all animals in the data set, even those animals that have not been genotyped.

Researchers at Iowa State, lead by Rohan Fernando, have created what they call a hybrid model which estimates marker effects for genotyped animals and breeding values for animals without genotypes.

With Simmental data, Bruce Golden see extra improvement in the precision when using a single-step Bayesian regression with different weights for DNA variants. This approach nearly cuts in half the uncertainty of EPDs compared with pedigree estimates.

This approach appears to be a major step forward in the precision of EPD estimation.


Monday, October 24, 2016

eBEEF Monday: How to Get Started with DNA Testing

This fact sheet goes through the fundamentals of how and when producers might use DNA testing in beef cattle production.  It covers the different types of tests that are available, how to submit samples and to whom, and what to do with the results.

For more information, see the eBEEF.org factsheet "How to Get Started with DNA Testing".

Monday, October 17, 2016

eBEEF Monday: Recent Developments in Genetic Evaluations and Genomic Testing

The application of genomics to improve the accuracy of EPDs is a rapidly developing field. There are ongoing improvements in genotyping and sequencing technologies, statistical methods to increase the correlation between genomic predictions and true genetic merit, and the computing systems to handle the large datasets associated with animal breeding. One thing still remains true in the genomic age and that is the need to collect accurate phenotypic records. It is essential to ensure performance data, pedigree, and DNA information are recorded and reported accurately. Genomic predictions will only be as reliable as the data upon which they are based.  Although it might seem like the genomics era could signal the end of performance recording, the opposite is true. Now more than ever, it is important that producers accurately report data, and ensure that animals which are genotyped are correctly identified so that their information can contribute towards improving the accuracy of the genomic predictions of the future.

For more information see the eBEEF.org factsheet "Recent Developments in Genetic Evaluations and Genomic Testing".

eBEEF Monday: Recent Developments in Genetic Evaluations and Genomic Testing

The application of genomics to improve the accuracy of EPDs is a rapidly developing field. There are ongoing improvements in genotyping and sequencing technologies, statistical methods to increase the correlation between genomic predictions and true genetic merit, and the computing systems to handle the large datasets associated with animal breeding. One thing still remains true in the genomic age and that is the need to collect accurate phenotypic records. It is essential to ensure performance data, pedigree, and DNA information are recorded and reported accurately. Genomic predictions will only be as reliable as the data upon which they are based.  Although it might seem like the genomics era could signal the end of performance recording, the opposite is true. Now more than ever, it is important that producers accurately report data, and ensure that animals which are genotyped are correctly identified so that their information can contribute towards improving the accuracy of the genomic predictions of the future.

For more information see the eBEEF.org factsheet "Recent Developments in Genetic Evaluations and Genomic Testing".

eBEEF Monday: Recent Developments in Genetic Evaluations and Genomic Testing

The application of genomics to improve the accuracy of EPDs is a rapidly developing field. There are ongoing improvements in genotyping and sequencing technologies, statistical methods to increase the correlation between genomic predictions and true genetic merit, and the computing systems to handle the large datasets associated with animal breeding. One thing still remains true in the genomic age and that is the need to collect accurate phenotypic records. It is essential to ensure performance data, pedigree, and DNA information are recorded and reported accurately. Genomic predictions will only be as reliable as the data upon which they are based.  Although it might seem like the genomics era could signal the end of performance recording, the opposite is true. Now more than ever, it is important that producers accurately report data, and ensure that animals which are genotyped are correctly identified so that their information can contribute towards improving the accuracy of the genomic predictions of the future.

For more information see the eBEEF.org factsheet "Recent Developments in Genetic Evaluations and Genomic Testing".

Monday, October 10, 2016

eBEEF Monday: Commercial Replacement Heifer Selection

Heifer selection is an important aspect of commercial beef operations, but unlike bull selection must be done without the aid of Expected Progeny Differences. This factsheet discusses considerations when making heifer selections, including available genomics tools and the importance of sire selection when replacement heifers are to be retained.

For more information, see the eBEEF.org factsheet.

Wednesday, October 5, 2016

NCBEC Brown Bagger: Potential impacts of functional variants on national cattle evaluation

Larry Kuehn
USDA-MARC

When we go from less than a thousand animals to several thousands of animals, genomic predictions can explain about 50% of the genetic variance for important traits. Genomic prediction is working and providing tremendous benefits to seedstock and commercial producers.

But, we still struggle with genomic predictions with very little data recording and genomic predictions that work well across breeds.

Two methods are used to use genomics in national cattle evaluation. With the genomic pedigree method you track genetic effects more accurately than with pedigree data. With the second method you are relying on linkage on chromosomes between the DNA markers and the variants responsible for the differences (causal mutations).

The linkage signal between DNA markers and causal variants breaks down over generations due to recombination (switching) between paternal and maternal chromosomes. Because this linkage breaks down over time is part of the reason genomic predicitons don't work well across breeds.

When we train a prediction in Angus and use the predictions in Red Angus, the predictive ability of the genomic prediction goes down significantly.

Table 1. Correlations from genomic predictions trained in Angus and used in Angus or Red Angus.
Breed Weaning Weight Yearling Weight
Train in Angus, Predict in Angus 0.36 0.51
Train in Angus, Predict in Red Angus 0.16 0.08

As we have additional whole genome (entire DNA) sequencing data, we will discover more DNA variants that affect the composition (sequence) or length of proteins. Many of these broken genes will likely affect fertility.

As we discover variants that appear to affect the function of proteins or are causal variants, we will need to use different methods to fully utilize this information. These DNA variants will need to be weighted differently in the genomic prediction, or genomic predictions will need to fit multiple classes of variants.

USDA has started a selection experiment in which they are selecting against variants that cause the protein coded by the gene not to function properly. The first calves from this experiment will be born in the spring.

Kuehn highlighted several research needs including:

  • Continued annotation (identification of genes and regulatory elements) of the reference genome
  • New sequence assemblies
  • Improvement of functional variant panels (DNA tests)
  • Improved imputation method and strategies (infer DNA variants not testing using the patterns of genotyped variants)
  • Continued detail oriented phenotyping
  • Improved methods to incorporate into national cattle evaluation
  • Gene expression difference (the amount of RNA produced by the same gene in different animals)
  • Evaluating cellular expression


Kuehn and coworkers believe that functional variants offers new opportunities for national cattle evaluation.

Monday, October 3, 2016

eBEEF.org Monday: The Genetics of Horned, Polled and Scurred Cattle

The condition of horned, polled or scurred in cattle has important economic and welfare considerations, but is poorly understood. This factsheet explores the genetic aspects of these conditions, their relationships with each other and how to manage them in your breeding program.

For more information see the factsheet "The Genetics of Horned, Polled and Scurred Cattle" on eBEEF.org.